Global Biodiversity Scenarios for the Year 2100

Author(s): Osvaldo E. Sala, F. Stuart Chapin III, Juan J. Armesto, Eric Berlow, Janine
Bloomfield, Rodolfo Dirzo, Elisabeth Huber-Sanwald, Laura F. Huenneke, Robert B.
Jackson, Ann Kinzig, Rik Leemans, David M. Lodge, Harold A. Mooney, Martín
Oesterheld, N. LeRoy Poff, Martin T. Sykes, Brian H. Walker, Marilyn Walker and
Diana H. Wall
Source: Science , Mar. 10, 2000, New Series, Vol. 287, No. 5459 (Mar. 10, 2000), pp. 1770-
1774
Published by: American Association for the Advancement of Science

Stable URL: https://www.jstor.org/stable/3074591

REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/3074591?seq=1&cid=pdf-
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

American Association for the Advancement of Science is collaborating with JSTOR to digitize,
preserve and extend access to Science

This content downloaded from

137.132.123.69 on Tue, 18 Oct 2022 04:19:04 UTC
All use subject to https://about.jstor.org/terms
 Global Biodiversity Scenarios for the Year 2100
Osvaldo E. Sala,l* F. Stuart Chapin 111,2 Juan J. Armesto,4 Eric Berlow,5 Janine Bloomfield,6 Rodolfo Dirzo,7
Elisabeth Huber-Sanwald,8 Laura F. Huenneke,9 Robert B. Jackson,'0 Ann Kinzig,'1 Rik Leemans,12 David M. Lodge,13
Harold A. Mooney,14 Martin Oesterheld,1 N. LeRoy Poff,'5 Martin T. Sykes,17 Brian H. Walker,'18 Marilyn Walker,3
Diana H. Wall16

most important determinants of changes in

Scenarios of changes in biodiversity for the year 2100 can now be developed basedbiodiversity
on at the global scale: changes in
scenarios of changes in atmospheric carbon dioxide, climate, vegetation, and landland use use, atmospheric CO2 concentration, ni-
and the known sensitivity of biodiversity to these changes. This study identified a
trogen deposition and acid rain, climate, and
ranking of the importance of drivers of change, a ranking of the biomes with respect biotic exchanges (deliberate or accidental in-
to expected changes, and the major sources of uncertainties. For terrestrial ecosys- troduction of plants and animals to an eco-
tems, land-use change probably will have the largest effect, followed by climate system). Second, we calculated the expected
change, nitrogen deposition, biotic exchange, and elevated carbon dioxide concentra- change of these drivers in each biome. Third,
tion. For freshwater ecosystems, biotic exchange is much more important. Mediter- we estimated for each biome the impact that
ranean climate and grassland ecosystems likely will experience the greatest propor- a unit change in each driver has on biodiver-
tional change in biodiversity because of the substantial influence of all drivers of sity. Finally, we derived three scenarios of
biodiversity change. Northern temperate ecosystems are estimated to experiencefuture the biodiversity for each biome, relative to
least biodiversity change because major land-use change has already occurred. Plau-
its initial diversity, based on alternative as-
sible changes in biodiversity in other biomes depend on interactions among the causes
sumptions about interactions among the driv-
of biodiversity change. These interactions represent one of the largest uncertainties in
ers of biodiversity change. We assumed that
projections of future biodiversity change.
(i) there are no interactions among the vari-
ous causes of biodiversity change, (ii) there
are antagonistic interactions and biodiversity
G lobal biodiversity is changing at an we include all terrestrial and freshwater will respond only to the driver to which it is
unprecedented rate (1, 2) as a complex organisms-including plants, animals, andmost mi-sensitive, or (iii) there are synergistic
response to several human-induced crobes-at scales ranging from genetic diversi-interactions and biodiversity will respond
changes in the global environment (3). The ty within populations, to species diversity, multiplicatively
to to the drivers of biodiversity
magnitude of this change is so large (1) and community diversity across landscapes. change.
Our Because the nature of interactions
so strongly linked to ecosystem processes (4, definition excludes exotic organisms thatamong have causes of biodiversity change is poor-
5) and society's use of natural resources (6, been introduced and communities such asly known, we present all three alternatives as
agri-
7) that biodiversity change is now consid- cultural fields that are maintained by regularplausible scenarios of biodiversity change.
ered an important global change in its own human intervention. We do not consider marine
right (8). In our definition of biodiversity, systems in this study. Drivers of Change
International conventions seek to mini- We used a business-as-usual scenario gener-
1Department of Ecology and Instituto de Investiga-
mize changes in biodiversity, just as other ated by global models of climate (Had CM2),
clones Fisiol6gicas y Ecol6gicas vinculadas a la Agri- conventions seek to reduce the atmospheric vegetation (Biome3) (12), and land use [Al
cultura, Faculty of Agronomy, University of Buenos concentration of CO2 and chlorofluorocar- scenario of Image 2 (13)] to estimate the
Aires, Avenida San Martin 4453, Buenos Aires 1417,
bons. Scientists and policy-makers are famil-change in magnitude of the drivers of biodi-
Argentina. 2Institute of Arctic Biology, 3lnstitute of
iar with, and frequently use, scenarios of versity change for each biome between 1990
Northern Forest Cooperative Research, University of
Alaska, Fairbanks, AK 99775, USA. 4Facultad de Cien-change in climate or of concentrations of and the year 2100. Our 10 terrestrial biomes
cias, Universidad de Chile, Casilla 653, Santiago, Chile.greenhouse gases in projecting the future resulted from aggregating the original Bailey
5Department of Integrative Biology, University of Cal- state of the global environment (9). Although ecoregions (14). We ranked the projected
ifornia, Berkeley, CA 94720, USA. 6Environmental De-
biodiversity changes are just as important for changes in drivers as small (value of 1) to
fense Fund, 257 Park Avenue, New York, NY 10010,
USA. 7lnstituto de Ecologia, UNAM, Mexico 04510,the functioning of ecosystems and the well- large (value of 5). We used the A scenario of
Mexico. 8Lehrstuhl fur Grunlandlehre, Technische Uni- being of humans, there are currently no sce- the IMAGE model to estimate changes in
versitat Munchen, D85350, Germany. 9Department ofnarios for biodiversity comparable to those of land use for each biome (13). The IMAGE
Biology, New Mexico State University, Las Cruces, NM
climate and greenhouse gases. Previous exer- model projects that most land-use change will
88003, USA. 'Department of Botany, Duke Univer-
cises have assessed extinction threats as a continue to occur in the tropical forests and in
sity, Durham, NC 27708, USA. 11Department of Biol-
function of human land use at the globalthe
ogy, Arizona State University, Tempe, AZ, 85287, temperate forests of South America and
and
USA. 12National Institute for Public Health & the regional levels (10, 11). that the least land-use change will occur in
Environment, Bilthoven, Netherlands. '3Department the arctic and alpine (where human popula-
of Biology, University of Notre Dame, Notre Dame, IN
Modeling Biodiversity Change tion density probably will remain low) and in
46556-0369 USA. 14Department of Biological Sci-
ences, Stanford University, Stanford, CA 94305, USA. We developed global scenarios of biodiver- northern temperate forests (where reforesta-
'5Department of Biology and 16Natural Resource sity change in 10 terrestrial biomes and tion
in is expected to exceed deforestation, also
Ecology Laboratory, Colorado State University, Fort freshwater ecosystems for the year 2100 causing small negative effects on biodiver-
Collins, CO 80523, USA. 17Ekologihuset, Lund Univer-
based on global scenarios of changes in en- sity) (Table 1). The extent of habitat modifi-
sity, 22362 Lund, Sweden. '8Division of Wildlife and
Ecology, Commonwealth Scientific and Industrial Re- vironment and land use and the understand- cation is projected to be modest in desert and
search Organization, Canberra, Australia. boreal forest and intermediate in savannas,
ing by ecological experts of the sensitivity of
*To whom correspondence should be addressed. E- biodiversity in each terrestrial biome to these grasslands, and Mediterranean ecosystems.
mail: [email protected] global changes. First, we identified the five Atmospheric CO2 mixes globally within a

1770 10 MARCH 2000 VOL 287 SCIENCE www.sciencemag.org

This content downloaded from

137.132.123.69 on Tue, 18 Oct 2022 04:19:04 UTC
All use subject to https://about.jstor.org/terms
 SCIENCE'S COMPASS

year (-15), so we assumed that largest impact factor to temperate forests,diversity, minimize the probability of suc-
all biomes
would experience the same changeborealin CO2 arctic, and alpine. Biodiversitycessful establishment by invaders in undis-
forests,
in deserts
concentration. Nitrogen deposition and tropical forests may respondturbed communities (25). Conversely, we ex-
is largest
least cities
in the northern temperate zone near to nitrogen
and deposition because plant pect the greatest effect of biotic exchange in
is smallest in biomes such as the arctic and growth is strongly limited by water and phos- biomes such as Mediterranean and southern
southern temperate forests, which generally phorus, respectively (23). Grasslands, savan-temperate forests that have long been isolated
are distant from sources of pollution. Othernas, and Mediterranean systems received in-and exhibit extensive convergent evolution
biomes are intermediate, with regional varia-termediate impact factors because nitrogen(26). Other biomes are intermediate in their
tion in deposition generally associated withand other factors limit plant growth. connectedness. There is wide variation within
cities or industrial point sources. Climate is A given change in climate is expected tomost biomes in the successful establishment
expected to warm most dramatically at highhave the largest proportional effect on biodiver-of biotic introductions, depending on the
latitudes (arctic and boreal zones), to changesity in those biomes characteristic of extremeoriginal diversity and isolation from similar
least in the tropics, and to show intermedi-climates, although biodiversity in all biomeshabitats. For example, islands typically have
ate changes in other biomes (9) (Table 1). likely will be sensitive to climate. Small chang- low diversity and are more prone to biotic
Changes in precipitation are uncertain and are es in temperature or precipitation in arctic, al- invasions (27).
difficult to generalize at the biome level. The pine, desert, and boreal forest will result in large When averaged across biomes, land-use
pattern of biotic exchange reflects the pattern changes in species composition and biodiver-change is the driver that is expected to have
of human activity. Remote areas with little sity. Similarly, we assume that biomes wherethe largest global impact on biodiversity by
human intervention receive fewer exotic spe- climate less strongly limits the activity of or-the year 2100 (Fig. 1), mostly because of
cies than areas that are in the middle of trade ganisms will experience changes in the distri-its devastating effects on habitat availability
routes or that host intense human activity (16). bution of organisms, but the overall effect onand consequent species extinctions. Climate
The second step of our exercise was to proportional change in diversity may be less change will be the second most important
evaluate, for each biome, the impact that a pronounced than in extreme environments. driver of biodiversity change, mostly as a
unit change in each driver has on biodiversity Biotic introductions (that is, successful result of the expected warming at high lati-
independently of the expected magnitude of establishment of exotic species) vary accord- tudes. Changes in atmospheric CO2, biotic
change in the driver (Table 2). Land-use ing to environmental conditions and biogeo-exchange, and nitrogen deposition also will
change is the most severe driver of changes in graphic considerations. Invasions have oc- have substantial effects on future biodiver-
biodiversity (17). For example, conversion of curred least frequently in arctic and alpinesity, with the relative importance being re-
temperate grasslands into croplands or tropi- ecosystems, because of their severe environ- gionally variable. Variability among biomes
cal forests into grasslands results in local ment (24) and the broad longitudinal distri-of the impact of the different drivers is max-
extinction of most plant species and the as- bution of much of the high-latitude flora andimal for land use, reflecting the broad range
sociated animals whose habitat is largely de- fauna. In the tropics, we also expect a smallof expected changes in this driver and the
termined by plant species composition. Be- proportional change in the diversity of intactlarge sensitivity of all biomes to land-use
low-ground organisms are also affected most ecosystems because of the high initial diver- change. In contrast, atmospheric CO2 showed
severely by land-use change (18). We as- sity and because abiotic and biotic factors the smallest variability of the three drivers
sumed no differences among biomes in the characteristic of this biome, including its high because CO2 is well mixed in the atmosphere
response to a unit change in land use and we
assigned land use the maximum impact factor Table 1. Expected changes for the year 2100 in the five major drivers of biodiversity change (land use,
because of the consistently large effect of land- atmospheric composition CO2, nitrogen deposition, climate, and biotic exchange) for the principal
use change on biodiversity. The increase in terrestrial biomes of the Earth (arctic tundra, alpine tundra, boreal forest, grasslands, savannas, Medi-
terranean ecosystems, deserts, northern temperate forests, southern temperate forests, and tropical
atmospheric CO2 is expected to have the largest
forests).
effect on biodiversity in those biomes where
plant growth is most limited by water availabil-
ity and where there is a mixture of C3 and C4 Arctic Alpine Boreal Med Desert N Tropic
land vanna temp temp
species because of known species differences in
the effect of CO2 on water-use efficiency (19, Land use 1.0 1.0 2.0 3.0 3.0 3.0 2.0 1.0 4.0 5.0
20). For example, changes in atmospheric CO2 Climate 5.0 3.0 4.0 2.0 2.0 2.0 2.0 2.0 2.0 1.0
Nitrogen deposition 1.0 3.0 3.0 3.0 2.0 3.0 2.0 5.0 1.0 2.0
may change the competitive balance between
Biotic exchange 1.0 1.0 2.0 3.0 3.0 5.0 3.0 3.0 2.0 2.0
species that differ in rooting depth, photosyn-
Atmospheric CO2 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5
thetic pathway, or woodiness as well as associ-
ated below-ground organisms (21). Therefore,
we assigned the maximum impact factor of Table 2. Impact of a large change in each driver on the biodiversity
elevated CO2 to grasslands and savannas, change of the driver was defined for land use as conversion of 50%
which are water-limited biomes with a mixture as a 2.5-fold increase in elevated CO2 as projected by 2100, for nitro
of contrasting plant functional types. Based on for climate as a 4?C change or 30% change in precipitation, and for
the same reasoning, we assigned the smallest new plant or animal species by 2100. Estimates vary from low (1
global scenarios of the physical environment and knowledge from
impact factors to arctic, alpine, boreal forest,
tropical forest, and freshwater ecosystems.
Increased nitrogen deposition should have Arctic Alpine Boreal land vann Med Desert Tropic
land vanna temp temp
the largest impact on biodiversity in those
biomes that are most nitrogen-limited primar- Land use 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.0
Climate 4.0 4.0 3.5 3.0 3.0 3.0 4.0 2.0 2.0 3.0
ily by giving a competitive advantage to plant
Nitrogen deposition 3.0 3.0 3.0 2.0 2.0 2.0 1.0 3.0 3.0 1.0
species with high maximum growth rates,
Biotic exchange 1.0 1.0 1.0 2.0 2.0 3.0 2.0 1.5 3.0 1.5
which then exclude the slower growing spe-
Atmospheric CO2 1.0 1.0 1.0 3.0 3.0 2.0 2.0 1.5 1.5 1.0
cies (22). Consequently, we assigned the

www.sciencemag.org SCIENCE VOL 287 10 MARCH 2000 1771

This content downloaded from

137.132.123.69 on Tue, 18 Oct 2022 04:19:04 UTC
All use subject to https://about.jstor.org/terms
 SCIENCE'S COMPASS

and the range of ecological responses systemsisarequite

also influenced largely by a sin- nitrogen deposition) and its interactions with
narrow. The other drivers have intermediate gle factor (climate change). In contrast, Med- climate change, land use, and stratospheric
variability. In this global analysis, we consid- iterranean ecosystems, savannas, and grass- ozone depletion are large especially for bo-
er only proportional changes in diversity and lands are substantially affected by most driv- real lakes (32). Recent analyses suggest that,
give no weighting to the area, initial species ers. Finally, biomes such as the northern as a result of all these impacts, global fresh-
diversity, or economic value of biomes. temperate forests and deserts show contribu- water biodiversity is declining at far greater
We performed a simple sensitivity analy- tions by all the drivers but most of them are rates than is true for even the most affected
sis (28) of our model by independently in- moderate. terrestrial ecosystems (33).
creasing and then decreasing by 10% the Freshwater ecosystems show substantial For streams, variation in expected impact
expected change of each driver. The ranking impacts from land use, biotic exchange, and exists along a latitudinal gradient from trop-
of drivers (Fig. 1) was not altered in any of climate (Fig. 2). Land use is expected to have ical to temperate to high latitude/altitude re-
the trials in which we increased the expected especially large effects because humans live gions. In tropical streams, land use is expect-
change of each driver. When we decreased disproportionately near waterways and exten- ed to have the greatest effect, with climate
each driver by 10%, in only one trial was the sively modify riparian zones even in terres- and biotic exchange being minimal. In tem-
ranking altered: nitrogen deposition switched trial biomes that otherwise are sparsely pop- perate streams, biodiversity will be similarly
positions with biotic exchange. These results ulated. This leads to many changes within the affected by both land-use change (34) and
suggest that modifications of the parameters waterways, including increased inputs of nu- biotic exchange (35), which reaches its max-
in +10% will not modify the result of this trients, sediments, and contaminants (29). In imum impact value in this region. In high
exercise with regard to the ranking of drivers. addition, humans use waterways as transpor- latitude/altitude streams, climate change is
The ranking of relative impact of global- tation corridors, sewage disposal sites, and the dominant driver and it is expected to
change drivers on biodiversity is relatively water sources, so that much of Earth's acces- cause the greatest change in biodiversity
insensitive to small changes in drivers be- sible freshwaters are already coopted by hu- (36), with land use and biotic exchange being
cause our model assumes no nonlinearities or mans (30). Biotic exchange, in particular, is minimal. Biodiversity in streams and rivers
bifurcations. Our scenarios result from mul- relatively more important for aquatic (espe- generally is more sensitive to climate than in
tiplication of the expected changes and bi- cially lakes) than for terrestrial ecosystems lakes because streams have greater respon-
ome-sensitivity matrices (Tables 1 and 2) and because of both extensive intentional (for siveness to runoff; generally, it is less sensi-
the linear combination of their product to example, fish stocking) and unintentional (for tive to biotic exchange because streams are
construct the ranking of drivers and the dif- example, ballast water releases) releases of physically harsh and more dynamic tempo-
ferent scenarios. organisms (31). Carbon dioxide and nitrogen rally (37).
deposition generally had less impact on lakes To estimate the total change in biodiver-
Variation Across Biomes
and streams than on terrestrial ecosystems, sity for each terrestrial biome, we provide
There are large differences among biomes but inacidic deposition (partly attributable to three alternative scenarios of biodiversity
the causes of future change in biodiversity based on the assumptions of no interactions,
(Fig. 2). Biomes such as tropical and southern1. Arctic 1 antagonistic interactions, or synergistic inter-
Mediterranean

temperate forest show large changes, mostly 0.6: actions among causes of biodiversity change.
due to changes in land use with relatively
0.2- In all scenarios, we project that grasslands
small effects due to other drivers. Arctic eco- and Mediterranean ecosystems will experi-
Alpine Desert
1-
Alpine 1 i Desert ence large biodiversity loss because of their
0.6- 0.6
1.2 sensitivity to all drivers of biodiversity
I change, particularly land-use change (Figs. 2
U 1.0 o 1- Boreal 1 N Temp and 3). We did not generate these scenarios
?0.6- 06
'o 0.8 Table 3. Expected changes for the year 2100 in
the major drivers of biodiversity change for lakes
0.6 1- Grassland 11 STemp and streams.
L 0.6-2 0.6

*0.2- _t .............. P~~/ 0.2 Lakes Streams

> 0.4
'>0
Land use 4.0 5.0
0.2
I 0.2 0.6n _ 0.6 Climate 3.0 4.0
0.2 [::-::- 0.2
Nitrogen deposition 2.0 2.0
0* Biotic exchange 5.0 3.5
Atmospheric CO2 2.5 2.5

c' ^o'" 0.2 0.2

1 2 3 4 5 1 2 3 4 5 Table 4. Impact of a large change i

the biodiversity of each major fre
Fig. of
Fig. 1. Relative effect of major drivers 2.chang-
Effect of eachtem driver
type. Methodson biodiversi
and assumpti
es on biodiversity. Expected biodiversity
change change
for each terrestrial
as in Tables 1 and biome
2. and freshw
for each biome for the year 2100 terwas calculat-
ecosystem type calculated as the product
ed as the product of the expectedthe change
expectedin change of each driver times it
Lakes Streams
drivers times the impact of each driver on
impact for each terrestrial biome or freshwat
biodiversity for each biome. Values are averag-
ecosystem. Expected Land changes useand5.0 impacts
5.0 ar
es of the estimates for each biome and they are
specific to each biome or ecosystem
Climate 3.0 4.0 type a
made relative to the maximum arechange, which
presented in Tables 1 to 4. Values are Nitrogen deposition 2.0 1.0
resulted from change in land use. Thin
relative bars
to the are possible value. Bars: Biotic
maximum 1, exchange 5.0 3.0
land use; 2, climate;
standard errors and represent variability among 3, nitrogen deposition; Atmospheric
4, CO2 1.0 1.0
biomes. biotic exchange; 5, atmospheric CO2.

1772 10 MARCH 2000 VOL 287 SCIENCE www.sciencemag.or

This content downloaded from

137.132.123.69 on Tue, 18 Oct 2022 04:19:04 UTC
All use subject to https://about.jstor.org/terms
 SCIENCE'S COMPASS

for freshwater ecosystems becausecrop species,

they further
are damage to biodiversitythe driver with the greatest impact. The
by otherbiomes.
distributed throughout all terrestrial drivers may not be possible. In suchstrength of interactions among drivers in their
Projected biodiversity changes incases, biodiversity
other ter- change responds only toeffects on biodiversity is virtually unknown.
restrial biomes differ dramatically among our
three scenarios.
If we assume that diversity will respond to
global changes, without any interaction
among these drivers of change, we project
that Mediterranean and grassland ecosystems
will be most sensitive to change (Figs. 2 and
3A). In contrast, arctic, alpine, and desert
ecosystems will show only moderate changes
in biodiversity for reasons that are specific to
each biome. The range of changes among
biomes projected by this scenario is relatively
small, with the changes in all biomes being
within 60% of the maximum change.
If we assume that diversity in each biome
will be determined only by the factor that has
the greatest impact on diversity, then we project
that tropical and southern temperate forests will
B "^ '
experience substantial changes in diversity due
to land-use change and the arctic will experi-
ence change due to climate change (Figs. 2 and _ ~.. L .^. '-
3B). In this scenario, deserts and alpine will tar s ote
show the fewest diversity changes, because
there is no single driver to which biodiversity in
these biomes is extremely sensitive.
If there are synergistic interactions among
all causes of biodiversity change, we project
that Mediterranean and grassland ecosystems
will experience the greatest biodiversity change
because diversity in these biomes is sensitive to
all global-change drivers (Figs. 2 and 3C). In
this scenario, tropical forest, arctic, and alpine .

ecosystems will show the fewest biodiversity

changes, because there are several drivers of
change to which these biomes are relatively
insensitive. In contrast to the no-interaction sce-
nario, in this case the range of expected change
is quite broad, encompassing two orders of
magnitude, because of the effect of synergistic
interactions on amplifying differences among
biomes.

Uncertainties
This analysis highlights the sensitivity of
biodiversity change to our assumptions about
interactions among causes of biodiversity
change. Which assumptions are most plausi-
ble? There is clear evidence for nonlinearities
and synergistic interactions among many of
the global change drivers. Invasions of exotic
species are promoted by human disturbance Fig. 3. Maps of three scenarios of the expected change in biodiversity for the year 2100. (A) There are
and changes in climate variability (interaction no interactions among drivers of biodiversity change; consequently, total change is calculated as the
sum of the effects of each driver, which in turn result from multiplying the expected change in the driver
of biotic exchange, land-use change, and cli-
for a particular biome (Table 1) times the impact of the driver, which is also a biome-specific
mate change). Elevated CO2 has the greatest
characteristic (Table 2). (B) Total biodiversity change equals the change resulting from the driver that
effect on species composition in the presence is expected to have the largest effect and is calculated as the maximum of the effects of all the drivers.
of nitrogen deposition (interaction of CO2 (C) Interactions among the drivers are synergistic; consequently, total change is calculated as the
and nitrogen deposition). Synergistic interac- product of the changes resulting from the action of each driver. Different colors represent expected
tions may decrease in importance at extreme change in biodiversity from moderate to maximum for the different biomes of the world ranked
according to total expected change. Numbers in parentheses represent total change in biodiversity
values of individual drivers of biodiversity
relative to the maximum value projected for each scenario. Biomes are Mediterranean ecosystems
change. For example, where land use has
(MED), grasslands (GRAS), savannas (SAV), boreal forest (BOR), southern temperate forest (S.TEMP),
been severe and extensive such as in forest
tropical forest (TROP), northern temperate forest (N.TEMP), arctic ecosystems (ARCT), and desert
clearing followed by seeding of an exotic(DESERT). Values for alpine, stream, and lake ecosystems are not shown.

www.sciencemag.org SCIENCE VOL 287 10 MARCH 2000 1773

This content downloaded from

137.132.123.69 on Tue, 18 Oct 2022 04:19:04 UTC
All use subject to https://about.jstor.org/terms
 SCIENCE'S COMPASS

We hypothesize that future changes in biodi-

Earth system. This analysis represents an at- 21. H. A. Mooney et al., in The Terrestrial Biosphere and
versity will be intermediate between
tempt toscenar-
develop future global biodiversity sce- Global Change: Implications for Natural and Managed
Ecosystems: A Synthesis of GCTE and Related Re-
ios that consider synergistic interactions or no of these scenarios to the
narios. Refinement
search, B. H. Walker, W. L. Steffen, J. Canadell, J. S. I.
interactions, but realistic projections
pointof
thatfuture
they are useful to policy-makers will Ingram, Eds. (Cambridge Univ. Press, Cambridge,
biodiversity change require improved
require under-
quantitative regional analyses and a 1999), pp. 141-189.
standing about interactions among drivers of 22. D. Tilman, Ecology. 74, 2179 (1993).
study of the interactions among factors to which
23. P. M. Vitousek, Ecology 65, 285 (1984).
biodiversity change. local biodiversity is most sensitive. Mitigation 24. W. D. Billings, BioScience 23, 697 (1973).
Other uncertainties in our analysis include
of the expected effects on biodiversity identi- 25. M. Rejmanek, in Biodiversity and Ecosystem Processes
the magnitude and regional variation instudy
fied in this theshould encompass both reduc- in Tropical Forests, G. H. Orians, R. Dirzo, J. H. Cush-

future changes in drivers, as thoroughly ana- man, Eds. (Springer-Verlag, Berlin, 1996), pp. 153-172.
tion of the rate of change of the drivers at the
26. H. A. Mooney and E. L. Dunn, Evolution 24, 292
lyzed by Intergovernmental Panel onscale
global Climate
and development of management (1970).
Change (38). This reflects future policies
practices specifically tailored for each region 27. P. M. Vitousek, L Loope, H. Adsersen, Eds., Islands:
governing (i) the intensity and aerial extent
according to its of
biological, social, and econom- Biological Diversity and Ecosystem Function (Springer-
Verlag, Berlin, 1995).
land-use change, (ii) the protection of biodi-
ic characteristics.
28. W. Scott Overton, in Ecosystem Modeling in Theory
versity per unit of land-use change, and (iii) and Practice: An Introduction with Case Studies,
changes in atmospheric composition. Uncer-
References and Notes C. A. S. Hall and J. W. Day, Eds. (Wiley, New York,
1. S. I. Pimm, G. J. Russell, J. L. Gittelman, T. M. Brooks, 1977), pp. 49-74.
tainties in future climate and vegetation re-
Science 269, 347 (1995). 29. Committee on Inland Aquatic Ecosystems, Water and
flect the same policy uncertainties 2. We (38).
define As a
change in biodiversity at the biome level Science Technology Board, Commission on Geo-
result of the large policy-related uncertainties as the changes in number and relative abundance sciences, of Environment, and Resources, Freshwater
species that occur naturally in that biome.
in drivers, we emphasize that we have pre- Ecosystems (National Academy Press, Washington,
3. P. M. Vitousek, Ecology 75, 1861 (1994). DC, 1996).
sented scenarios rather than predictions of
4. F. S. Chapin et al., Science 277, 500 (1997). 30. S. L. Postel, G. C. Daily, P. R. Ehrlich, Science 271, 785
biodiversity change. 5. D. Tilman et al., Science 277, 1300 (1997). (1996).
We expect large regional variation in 6. G. C. Daily, Ed., Nature's Services. Societal Depen- 31. D. M. Lodge et al., Aust. J. Ecol. 23, 53 (1998).
dence on Natural Ecosystems (Island Press, Washing-
biodiversity change within each biome. For ton, DC, 1997). 32. D. W. Schindler, BioScience 48, 157 (1998).
33. A. Ricciardi and J. Rasmussen, Conserv. Biol. 13, 1220
example, diversity in islands, lakes, and some7. R. Costanza et al., Nature 387, 253 (1997).
(1999).
streams is particularly vulnerable to biotic 8. B. H. Walker and W. Steffen, Eds., Global Change and
34. J. S. Harding, E. F. Benfield, P. V. Bolstad, G. S.
Terrestrial Ecosystems (Cambridge Univ. Press, Cam-
exchange because geographic isolation has Helfman, E. B. D. Jones, Proc. Natl. Acad. Sci. U.S.A.
bridge, 1996).
led to local adaptation and, in the case of 95, 14843 (1998).
9. A. Kattenberg et al., in Climate Change: The IPCC
35. B. D. Richter, D. P. Braun, M. A. Mendelson, L. L.
islands, often a low biodiversity (27, 31). Hot Scientific Assessment, J. T. Houghton et al., Eds. (Cam-
Master, Conserv. Biol. 11, 1081 (1997).
spots of diversity such as riparian corridors bridge Univ. Press, Cambridge, 1996), pp. 285. 36. M. W. Oswood, A. M. Milner, J. G. Irons, in Global
10. D. S. Wilcove, D. Rothstein, J. Dubow, A. Phillips, E.
and coastal land margins often coincide with Losos, BioScience 48, 607 (1998). Climate Change and Freshwater Ecosystems, P. Firth
hot spots of development, leading to large 11. T. D. Sisk, A. E. Launer, K. R. Switky, P. R. Ehrlich, and S. G. Fischer, Eds. (Springer-Verlag, New York,
BioScience 44, 592 (1994). 1992), pp. 192-210.
biodiversity loss. Within-biome variation in
12. A. Haxeltine and I. C. Prentice, Global Biogeochem. 37. N. L. Poff et al., BioScience 47, 769 (1997).
climate may cause the greatest biodiversity Cycles 10, 693 (1996). 38. T. Houghton et al., Eds., Climate Change 1995: The
change near climatically determined bound-13. J. Alcamo, Image 2: Integrated Modeling of Global Science of Climate Change (Cambridge Univ. Press,
Climate Change (Kluwer Academic, Dordrecht, Neth- Cambridge, 1996).
aries of organism distribution. Other specific
erlands, 1994). 39. F. S. Chapin III, 0. E. Sala, E. Huber-Sanwald, Eds.,
local patterns of biodiversity change are less 14. R. G. Bailey, Ecoregions: The Ecosystem Geography of Future Scenarios of Global Biodiversity (Springer-Ver-
predictable than general trends at larger the Oceans and Continents (Springer-Verlag, New lag, New York, in press).
scales and may reflect interactions among York, 1998). 40. Supported by University of California, Santa Barbara,
15. I. Y. Fung, C. J. Tucker, K. C. Prentice, . Geophys. Res. National Center for Ecological Analysis and Synthe-
drivers of biodiversity change that are locally 92D, 2999 (1987). sis, InterAmerican Institute for Global Change Re-
important or a consequence of local "surpris- 16. J. A. Drake et al., Biological Invasions: A Global Per- search, National Science Foundation OCE-9634876,
es." An initial analysis of the causes of re- spective (Wiley, Chichester, UK, 1989), vol. 37. Consejo Superior de Investigaciones Cientificas y
17. 0. E. Sala, in Global Biodiversity Assessment, Section Tecnicas, University of Buenos Aires, and Fondo para
gional variation in diversity loss within each 5, H. A. Mooney, J. Lubchenco, R. Dirzo, O. E. Sala, Eds. la Investigaci6n Cientifica y Tecnolbgica de la Agen-
biome is being published separately (39). (Cambridge Univ. Press, Cambridge, 1995). cia Nacional de Promocibn Cientifica y Tecnolbgica.
Biodiversity in all biomes is sensitive to 18. J. M. Anderson, in Global Biodiversity Assessment: This exercise stems from an activity of the Global
Section 6, H. A. Mooney, J. Lubchenco, R. Dirzo, O. E. Change and Terrestrial Ecosystems (GCTE) core
global changes in environment and land use and
Sala, Eds. (Cambridge Univ. Press, Cambridge, 1995), project of the International Geosphere-Biosphere
realistic projections of biodiversity change will pp. 406-412. Programme (IGBP). We thank A. T; Austin, D. Tilman,
require an integrated effort by climatologists, 19. H. A. Mooney, B. G. Drake, R. J. Luxmoore, W. C. and W. Reid for their useful suggestions. A. E. Sala
Oechel, L. F. Pitelka, BioScience 41, 96 (1991). and J. P. Guerschman provided valuable assistance.
ecologists, social scientists, and policy makers 20. R. B. Jackson, 0. E. Sala, C. B. Field, H. A. Mooney,
to improve scenarios of future changes in the Oecologia 98, 257 (1994). 23 August 1999; accepted 7 December 1999

Enhance your AAAS membership with the All the information you need.....in one convenient location.
Science Online advantage. . -........- ..... . Visit Science Online at http://www.scienceonline.org,
* Full text Science-research papers and news articles ?* Research Alerts-sends you an e-mail alert every time call 202-326-6417, or e-mail [email protected]
with hyperlinks from citations to related abstracts in a Science research report comes out in the discipline, or by for more information.
other journals before you receive 'Science in the mail. a specific author, citation, or keyword of your choice.

* ScienceNOW-succinct, daily briefings, of the hottest e Science's Professional Network-lists hun- AAAS is also proud to announce site-wide institutional

scientific, medical, and technological news. dreds of job openings and funding sources worldwide subscriptions to Science Online. Contactyour subscription
* Science's Next Wave-career advice, topical that are quickly and easily searchable by discipline, posi-
forums, discussion groups, and expanded news written tion, organization, and region.
by today's brightest young scientists across the world. * Electronic Marketplace-provides new product

0 ^C! ^^.^^, * &.B l gB ginformation from the world's leading science manufac- AMERICAN ASSOCIATION FOR THE
SciCenic nO N L N E turers and suppliers, all at a click of your mouse. ADVANCEMENT OF SCIENCE

1774 10 MARCH 2000 VOL 287 SCIENCE www.sciencemag.org

This content downloaded from

137.132.123.69 on Tue, 18 Oct 2022 04:19:04 UTC
All use subject to https://about.jstor.org/terms